Road networks are critical infrastructures underpinning intelligent transportation systems and their related applications. Effective representation learning of road networks remains challenging due to the complex interplay between spatial structures and frequency characteristics in traffic patterns. Existing graph neural networks for modeling road networks predominantly fall into two paradigms: spatial-based methods that capture local topology but tend to over-smooth representations, and spectral-based methods that analyze global frequency components but often overlook localized variations. This spatial-spectral misalignment limits their modeling capacity for road networks exhibiting both coarse global trends and fine-grained local fluctuations. To bridge this gap, we propose HiFiNet, a novel hierarchical frequency-decomposition graph neural network that unifies spatial and spectral modeling. HiFiNet constructs a multi-level hierarchy of virtual nodes to enable localized frequency analysis, and employs a decomposition–updating–reconstruction framework with a topology-aware graph transformer to separately model and fuse low- and high-frequency signals. Theoretically justified and empirically validated on multiple real-world datasets across four downstream tasks, HiFiNet demonstrates superior performance and generalization ability in capturing effective road network representations.

Federated Edge Learning (FEL) has emerged as a promising approach for enabling edge devices to collaboratively train machine learning models while preserving data privacy. Despite its advantages, practical FEL deployment faces significant challenges related to device constraints and device-server interactions, necessitating heterogeneous, user-adaptive model training with limited and uncertain communication. While knowledge cache-driven federated learning offers a promising FEL solution for demanding edge environments, its logits-based interaction design provides poor richness of exchanged information for on-device model optimization. To tackle this issue, we introduce DistilCacheFL, a novel personalized FEL architecture that enhances the exchange of optimization insights while delivering state-of-the-art performance with efficient communication. DistilCacheFL incorporates the benefits of both dataset distillation and knowledge cache-driven federated learning by storing and organizing distilled data as knowledge in the server-side knowledge cache, allowing devices to periodically download and utilize personalized knowledge for local model optimization. Moreover, a device-centric cache sampling strategy is introduced to tailor transferred knowledge for individual devices within controlled communication bandwidth. Extensive experiments on five datasets covering image recognition, audio understanding, and mobile sensor data mining tasks demonstrate that (1) DistilCacheFL significantly outperforms state-of-the-art methods regardless of model structures, data distributions, and modalities. (2) DistilCacheFL can train splendid personalized on-device models with at least 28.6 improvement in communication efficiency.

Fine-tuning large language models is essential for task-specific adaptation, yet it remains computationally prohibitive. Parameter-Efficient Fine-Tuning (PEFT) methods have emerged as a solution, but current approaches typically ignore the distinct roles of model components and the heterogeneous importance across layers, thereby limiting adaptation efficiency. Motivated by the observation that Rotary Position Embeddings (RoPE) induce critical activations in the low-frequency dimensions of attention states, we propose RoPE-aware Selective Adaptation (RoSA), a novel PEFT framework that allocates trainable parameters in a more targeted and effective manner. RoSA comprises a RoPE-aware Attention Enhancement (RoAE) module, which selectively enhances the low-frequency components of RoPE-influenced attention states, and a Dynamic Layer Selection (DLS) strategy that adaptively identifies and updates the most critical layers based on LayerNorm gradient norms. By combining dimension-wise enhancement with layer-wise adaptation, RoSA achieves more targeted and efficient fine-tuning. Extensive experiments on fifteen commonsense and arithmetic benchmarks demonstrate that RoSA outperforms mainstream PEFT methods under comparable trainable parameters.

The growing demand for efficient and lightweight Retrieval-Augmented Generation (RAG) systems has highlighted significant challenges when deploying Small Language Models (SLMs) in existing RAG frameworks.Current approaches face severe performance degradation due to SLMs' limited semantic understanding and text processing capabilities, creating barriers for widespread adoption in resource-constrained scenarios.To address these fundamental limitations, we present Mini-RAG, a novel RAG system designed for simplicity and efficiency.MiniRAG introduces two key technical innovations: (1) a semanticaware heterogeneous graph indexing mechanism that combines text chunks and named entities in a unified structure, reducing reliance on complex semantic understanding, and (2) a lightweight topology-enhanced retrieval approach that leverages graph structures for efficient knowledge discovery without requiring advanced language capabilities.Our extensive experiments demonstrate that MiniRAG achieves comparable performance to LLMbased methods even when using SLMs while requiring only 25% of the storage space.Additionally, we contribute a comprehensive benchmark dataset for evaluating lightweight RAG systems under realistic on-device scenarios with complex queries.We fully open-source our implementation and datasets

Large language models (LLMs) have demonstrated remarkable progress in reasoning, often through supervised fine-tuning (SFT).However, SFT is resource-intensive, relying on large curated datasets, rejection-sampled demonstrations, and uniform optimization across all tokens-even though only a fraction carry meaningful learning value.In this work, we explore a counterintuitive idea: can smaller language models (SLMs) teach larger language models (LLMs) by revealing high-value reasoning moments that reflect the latter's unique strength?We propose LightReasoner 1 , a novel framework that leverages the behavioral divergence between a stronger expert model (LLM) and a weaker amateur model (SLM).LightReasoner operates in two stages: (1) a sampling stage that pinpoints critical reasoning moments and constructs supervision examples capturing the expert's advantage through expert-amateur contrast, and (2) a fine-tuning stage that aligns the expert model with these distilled examples, amplifying its reasoning strengths.Across seven benchmarks, LightReasoner improves accuracy by up to 28.1%, while reducing time consumption by 90%, sampled problems by 80%, and tuned token usage by 99%, all without relying on ground-truth labels.By turning weaker SLMs into effective teaching signals, LightReasoner offers a scalable and resource-efficient approach for advancing LLM reasoning.

Traffic state prediction based on spatiotemporal data has become a prominent focus in data-driven AI research. While significant progress has been made, most mainstream approaches assume uniform spatial and temporal correlations across conditions and use shared parameters for all scenarios. This simplification overlooks the complexity and heterogeneity inherent in human mobility patterns, often leading to suboptimal predictions. Recently, methods adopting the "decompose, then predict" (DTP) paradigm have gained traction. These methods break down data into smaller, manageable subcomponents, each predicted using dedicated parameters. Although effective in practice, DTP methods face unresolved theoretical questions: What type of decomposition truly makes subcomponents more manageable than the original data? To address this, we present an information theory-based analysis that derives sufficient conditions for a decomposition algorithm to reduce data-induced prediction errors. These conditions suggest that an effective algorithm should ensure decomposed components are as independent as possible, a principle we term the Component Independence Principle. Guided by this principle, we introduce the Theory-guided Graph Decomposition Learning (TGDL) framework, which decomposes graph-based multivariate time series data into approximately independent subgraph components that are easier to predict than the original data. Moreover, TGDL is a portable framework that can be integrated into any graph-based traffic prediction model to improve its predictive performance. Extensive experiments on four public datasets demonstrate the effectiveness of our approach. With a solid theoretical foundation, our TGDL enhances the performance of diverse traffic prediction models, yielding an average improvement of 19.37% across experiments.

Nowadays, navigation and ride-sharing apps have collected numerous images with spatio-temporal data. A core technology for analyzing such images, associated with spatio-temporal information, is Traffic Scene Understanding (TSU), which aims to provide a comprehensive description of the traffic scene. Unlike traditional spatio-temporal data analysis tasks, the dependence on both spatio-temporal and visual-textual data introduces distinct challenges to TSU task. However, recent research often treats TSU as a common image understanding task, ignoring the spatio-temporal information and overlooking the interrelations between different aspects of the traffic scene. To address these issues, we propose a novelSpatio-Temporal Enhanced Model based on CILP(ST-CLIP) for TSU. Our model uses the classic vision-language model, CLIP, as the backbone, and designs a Spatio-temporal Context Aware Multi-aspect Prompt (SCAMP) learning method to incorporate spatio-temporal information into TSU. The prompt learning method consists of two components: A dynamic spatio-temporal context representation module that extracts representation vectors of spatio-temporal data for each traffic scene image, and a bi-level ST-aware multi-aspect prompt learning module that integrates the ST-context representation vectors into word embeddings of prompts for the CLIP model. The second module also extracts low-level visual features and image-wise high-level semantic features to exploit interactive relations among different aspects of traffic scenes. To the best of our knowledge, this is the first attempt to integrate spatio-temporal information into vision-language models to facilitate TSU task. Experiments on two real-world datasets demonstrate superior performance in the complex scene understanding scenarios with a few-shot learning strategy.

POI representation learning plays a crucial role in handling tasks related to user mobility data. Recent studies have shown that enriching POI representations with multimodal information can significantly enhance their task performance. Previously, the textual information incorporated into POI representations typically involved only POI categories or check-in content, leading to relatively weak textual features in existing methods. In contrast, large language models (LLMs) trained on extensive text data have been found to possess rich textual knowledge. However leveraging such knowledge to enhance POI representation learning presents two key challenges: first, how to extract POI-related knowledge from LLMs effectively, and second, how to integrate the extracted information to enhance POI representations. To address these challenges, we propose POI-Enhancer, a portable framework that leverages LLMs to improve POI representations produced by classic POI learning models. We first design three specialized prompts to extract semantic information from LLMs efficiently. Then, the Dual Feature Alignment module enhances the quality of the extracted information, while the Semantic Feature Fusion module preserves its integrity. The Cross Attention Fusion module then fully adaptively integrates such high-quality information into POI representations and Multi-View Contrastive Learning further injects human-understandable semantic information into these representations. Extensive experiments on three real-world datasets demonstrate the effectiveness of our framework, showing significant improvements across all baseline representations.

Large Language Models (LLMs) possess remarkable generalization capabilities but struggle with multi-task adaptation, particularly in balancing knowledge retention with task-specific specialization. Conventional fine-tuning methods suffer from catastrophic forgetting and substantial resource consumption, while existing parameter-efficient methods perform suboptimally in complex multi-task scenarios. To address this, we propose Contextual Attention Modulation (CAM), a novel mechanism that dynamically modulates the representations of self-attention modules in LLMs. CAM enhances task-specific features while preserving general knowledge, thereby facilitating more effective and efficient adaptation. For effective multi-task adaptation, CAM is integrated into our Hybrid Contextual Attention Modulation (HyCAM) framework, which combines a shared, full-parameter CAM module with multiple specialized, lightweight CAM modules, enhanced by a dynamic routing strategy for adaptive knowledge fusion. Extensive experiments on heterogeneous tasks, including question answering, code generation, and logical reasoning, demonstrate that our approach significantly outperforms existing approaches, achieving an average performance improvement of 3.65%. The implemented code and data are available to ease reproducibility at https://github.com/Applied-Machine-Learning-Lab/HyCAM.

To effectively manage and utilize massive distributed data at the network edge, Federated Learning (FL) has emerged as a promising edge computing paradigm across data silos. However, FL still faces two challenges: system heterogeneity (i.e., the diversity of hardware resources across edge devices) and statistical heterogeneity (i.e., non-IID data). Although sparsification can extract diverse submodels for diverse clients, most sparse FL works either simply assign submodels with artificially-given rigid rules or prune partial parameters using heuristic strategies, resulting in inflexible sparsification and poor performance. In this work, we propose Learnable Personalized Sparsification for heterogeneous Federated learning (FedLPS), which achieves the learnable customization of heterogeneous sparse models with importance-associated patterns and adaptive ratios to simultaneously tackle system and statistical heterogeneity. Specifically, FedLPS learns the importance of model units on local data representation and further derives an importance-based sparse pattern with minimal heuristics to accurately extract personalized data features in non-IID settings. Furthermore, Prompt Upper Confidence Bound Variance (P-UCBV) is designed to adaptively determine sparse ratios by learning the superimposed effect of diverse device capabilities and non-IID data, aiming at resource self-adaptation with promising accuracy. Extensive experiments show that FedLPS outperforms status quo approaches in accuracy and training costs, which improves accuracy by 1.28% – 59.34% while reducing running time by more than 68.80%.

Spatiotemporal (ST) data analysis is a critical area of research in data engineering. Typical dynamic ST data includes trajectory data (representing individual-level mobility) and traffic state data (representing population-level mobility). Traditional studies often treat trajectory and traffic state data as distinct, independent modalities, each tailored to specific tasks within a single modality. However, real-world applications, such as navigation apps, require joint analysis of trajectory and traffic state data. Treating these data types as two separate domains can lead to suboptimal model performance. Although recent advances in ST data pre-training and ST foundation models aim to develop universal models for ST data analysis, most existing models are “multi-task, solo-data modality” (MTSM), meaning they can handle multiple tasks within either trajectory data or traffic state data, but not both simultaneously. To address this gap, this paper introduces BIGCity, a pioneer multi-task, multi-data modality (MTMD) model for ST data analysis. The model targets two key challenges in designing an MTMD ST model: (1) unifying the representations of different ST data modalities, and (2) unifying heterogeneous ST analysis tasks. To overcome the first challenge, BIGCity introduces a novel ST-unit that represents both trajectories and traffic states in a unified format. Additionally, for the second challenge, BIGCity adopts a tunable large model with ST task-oriented prompt, enabling it to perform a range of heterogeneous tasks without the need for fine-tuning. Extensive experiments on real-world datasets demonstrate that BIGCity achieves state-of-the-art performance across 8 tasks, outperforming 17 baselines. Our code are available at https://github.com/bigscity/BIGCity.

While deep learning has witnessed remarkable achievements in a wide range of applications, its substantial computational cost imposes limitations on application scenarios of neural networks. To alleviate this problem, low-rank compression is proposed as a class of efficient and hardware-friendly network compression methods, which reduce computation by replacing large matrices in neural networks with products of two small ones. In this paper, we implement low-rank networks by inserting a sufficiently narrow linear layer without bias between each of two adjacent nonlinear layers. We prove that low-rank Swish networks with a fixed depth are capable of approximating any function from the Hölder ball $\mathcal{C}^{\beta, R}([0,1]^d)$ within an arbitrarily small error where $\beta$ is the smooth parameter and $R$ is the radius. Our proposed constructive approximation ensures that the width of linear hidden layers required for approximation is no more than one-third of the width of nonlinear layers, which implies that the computational cost can be decreased by at least one-third compared with a network with the same depth and width of nonlinear layers but without narrow linear hidden layers. Our theoretical finding can offer a theoretical basis for low-rank compression from the perspective of universal approximation theory.

Recent years have witnessed the perfect encounter of deep learning and quantitative trading has achieved great success in stock investment. Numerous deep learning-based models have been developed for forecasting stock returns, leveraging the powerful representation capabilities of neural networks to identify patterns and factors influencing stock prices. These models can effectively capture general patterns in the market, such as stock price trends, volume-price relationships, and time variations. However, the impact of special irrationality factors -- such as market sentiment, speculative behavior, market manipulation, and psychological biases -- has not been fully considered in existing deep stock forecasting models due to their relative abstraction as well as lack of explicit labels and data description. To fill this gap, we propose UMI, a Universal multi-level Market Irrationality factor model to enhance stock return forecasting. The UMI model learns factors that can reflect irrational behaviors in market from both individual stock and overall market levels. For the stock-level, UMI construct an estimated rational price for each stock, which is cointegrated with the stock's actual price. The discrepancy between the actual and the rational prices serves as a factor to indicate stock-level irrational events. Additionally, we define market-level irrational behaviors as anomalous synchronous fluctuations of stocks within a market. Using two self-supervised representation learning tasks, i.e., sub-market comparative learning and market synchronism prediction, the UMI model incorporates market-level irrationalities into a market representation vector, which is then used as the market-level irrationality factor. We also developed a forecasting model that captures both temporal and relational dependencies among stocks, accommodating the UMI factors. Extensive experiments on U.S. and Chinese stock markets with competitive baselines demonstrate our model's effectiveness and the universality of our factors in improving various forecasting models. We provide our code at https://github.com/lIcIIl/UMI.

Electronic maps consist of diverse entities, such as points of interest (POIs), road segments, and land parcels, playing a vital role in applications like ITS and LBS. Map entity representation learning (MapRL) generates versatile and reusable data representations, providing essential tools for efficiently managing and utilizing map entity data. Despite the progress in MapRL, two key challenges constrain further development. First, existing research is fragmented, with models classified by the type of map entity, limiting the reusability of techniques across different tasks. Second, the lack of unified benchmarks makes systematic evaluation and comparison of models difficult. To address these challenges, we propose a novel taxonomy for MapRL that organizes models based on functional modules—such as encoders, pre-training tasks, and downstream tasks—rather than by entity type. Building on this taxonomy, we present a taxonomy-driven library, VecCity , which offers easy-to-use interfaces for encoding, pre-training, fine-tuning, and evaluation. The library integrates datasets from nine cities and reproduces 21 mainstream MapRL models, establishing the first standardized benchmarks for the field. VecCity also allows users to modify and extend models through modular components, facilitating seamless experimentation. Our comprehensive experiments cover multiple types of map entities and evaluate 21 VecCity pre-built models across various downstream tasks. Experimental results demonstrate the effectiveness of VecCity in streamlining model development and provide insights into the impact of various components on performance. By promoting modular design and reusability, VecCity offers a unified framework to advance research and innovation in MapRL. The code is available at https://github.com/Bigscity-VecCity/VecCity.

For distributed network traffic prediction with data localization and privacy protection, Federated Learning (FL) enables collaborative training without raw data exchange across Base Stations (BSs). Nevertheless, traffic across BSs exhibit inherently heterogeneous trend burst and smooth fluctuation properties, but existing FL methods model single-scale series from only one view, which cannot simultaneously capture diverse trend and fluctuation properties, especially distinct burst distributions. In this paper, we proposePersonalized Federated Forecasting with Multi-property Self-fusion (P2FMS), which can represent multi-scale traffic properties from different views. With precise multi-property representations, a fusion-level prediction decision is learned for each client in a personalized manner to promptly sense traffic bursts and improve forecasting performance in non-IID settings. Specifically, P2FMS decomposes the traffic series into distinct time scales, based on which, we effectively extract closeness, period, and trend properties from different views. The closeness and period are embedded through global-view representations with spatial correlations, while non-stationary trends are individually fitted from the client-side view. Furthermore, a personalized combiner is designed to accurately quantify the proportion of general fluctuation raws (i.e., closeness and period) and specific trend property in predictions, which enables multi-property self-fusion for each client to accommodate heterogeneous traffic patterns and enhance prediction accuracy. Besides, an alternant training mechanism is introduced to optimize property representation and fusion modules with the convergence guarantee. Extensive experiments on real-world datasets show that P2FMS outperforms status quo methods in both prediction performance and convergence time.

Significance testing aims to determine whether a proposition about the population distribution is the truth or not given observations. However, traditional significance testing often needs to derive the distribution of the testing statistic, failing to deal with complex nonlinear relationships. In this paper, we propose to conduct Full Bayesian Significance Testing for neural networks, called nFBST, to overcome the limitation in relationship characterization of traditional approaches. A Bayesian neural network is utilized to fit the nonlinear and multi-dimensional relationships with small errors and avoid hard theoretical derivation by computing the evidence value. Besides, nFBST can test not only global significance but also local and instance-wise significance, which previous testing methods don't focus on. Moreover, nFBST is a general framework that can be extended based on the measures selected, such as Grad-nFBST, LRP-nFBST, DeepLIFT-nFBST, LIME-nFBST. A range of experiments on both simulated and real data are conducted to show the advantages of our method.

Large language models (LLMs) are still struggling in aligning with human preference in complex tasks and scenarios.They are prone to overfit into the unexpected patterns or superficial styles in the training data.We conduct an empirical study that only selects the top-10% most updated parameters in LLMs for alignment training, and see improvements in the convergence process and final performance.It indicates the existence of redundant neurons in LLMs for alignment training.To reduce its influence, we propose a low-redundant alignment method named ALLO, focusing on optimizing the most related neurons with the most useful supervised signals.Concretely, we first identify the neurons that are related to the human preference data by a gradient-based strategy, then identify the alignment-related key tokens by reward models for computing loss.Besides, we also decompose the alignment process into the forgetting and learning stages, where we first forget the tokens with unaligned knowledge and then learn aligned knowledge, by updating different ratios of neurons, respectively.Experimental results on 10 datasets have shown the effectiveness of ALLO.Our code and data are available at https://github.com/RUCAIBox/ALLO.

Air quality prediction and modelling plays a pivotal role in public health and environment management, for individuals and authorities to make informed decisions. Although traditional data-driven models have shown promise in this domain, their long-term prediction accuracy can be limited, especially in scenarios with sparse or incomplete data and they often rely on black-box deep learning structures that lack solid physical foundation leading to reduced transparency and interpretability in predictions. To address these limitations, this paper presents a novel approach named Physics guided Neural Network for Air Quality Prediction (AirPhyNet). Specifically, we leverage two well-established physics principles of air particle movement (diffusion and advection) by representing them as differential equation networks. Then, we utilize a graph structure to integrate physics knowledge into a neural network architecture and exploit latent representations to capture spatio-temporal relationships within the air quality data. Experiments on two real-world benchmark datasets demonstrate that AirPhyNet outperforms state-of-the-art models for different testing scenarios including different lead time (24h, 48h, 72h), sparse data and sudden change prediction, achieving reduction in prediction errors up to 10\%. Moreover, a case study further validates that our model captures underlying physical processes of particle movement and generates accurate predictions with real physical meaning. The code is available at: https://github.com/kethmih/AirPhyNet

Recent years witness the increasing popularity of ride-on-demand (RoD) services such as Uber and Didi. Compared with traditional taxi, RoD service is more “data-driven” and adopts dynamic pricing to manipulate the supply and demand in real time. Dynamic price could be viewed as an accurate and quantitative indicator of the supply and demand, and could provide clues to drivers, passengers, and the service providers, possibly reshaping the ways in which some problems are solved. In this paper, we focus on the seeking route recommendation problem that aims at increasing driver revenue by recommending highly profitable seeking routes to drivers of vacant cars with the help of dynamic prices. We first justify our motivation by showing the importance of route recommendation and answering why it is necessary to consider dynamic prices, based on the analysis of real service data. We then design a dynamic price prediction model to generate the dynamic prices at any given time and location based on multi-source urban data. After that, a reinforcement learning model is adopted to perform seeking route recommendation based on predicted dynamic prices. We conduct extensive experiments in different spatio-temporal combinations and make comparisons with multiple baselines. Results first show that our dynamic price prediction model achieves an accuracy ranging from 83.82% to 90.67% under different settings. It also proves that considering the real-time predicted dynamic prices significantly increases driver revenue by, for example, 12% and 47.5% during weekday evening rush hours, than merely using the average prices or completely ignoring dynamic prices.

Abstract Public health surveillance based on data analytics plays a crucial role in detecting and responding to public health crises, such as infectious disease outbreaks. Previous information science research on the topic has focused on developing analytical algorithms and visualization tools. This study seeks to extend the research by investigating information practices in data analytics for public health surveillance. Through a case study of how data analytics was conducted for surveilling Influenza A and COVID‐19 outbreaks, both exploration information practices (i.e., probing, synthesizing, exchanging) and exploitation information practices (i.e., scavenging, adapting, outreaching) were identified and detailed. These findings enrich our empirical understanding of how data analytics can be implemented to support public health surveillance.

Robust prediction of citywide traffic flows at different time periods plays a crucial role in intelligent transportation systems. While previous work has made great efforts to model spatio-temporal correlations, existing methods still suffer from two key limitations: i) Most models collectively predict all regions' flows without accounting for spatial heterogeneity, i.e., different regions may have skewed traffic flow distributions. ii) These models fail to capture the temporal heterogeneity induced by time-varying traffic patterns, as they typically model temporal correlations with a shared parameterized space for all time periods. To tackle these challenges, we propose a novel Spatio-Temporal Self-Supervised Learning (ST-SSL) traffic prediction framework which enhances the traffic pattern representations to be reflective of both spatial and temporal heterogeneity, with auxiliary self-supervised learning paradigms. Specifically, our ST-SSL is built over an integrated module with temporal and spatial convolutions for encoding the information across space and time. To achieve the adaptive spatio-temporal self-supervised learning, our ST-SSL first performs the adaptive augmentation over the traffic flow graph data at both attribute- and structure-levels. On top of the augmented traffic graph, two SSL auxiliary tasks are constructed to supplement the main traffic prediction task with spatial and temporal heterogeneity-aware augmentation. Experiments on four benchmark datasets demonstrate that ST-SSL consistently outperforms various state-of-the-art baselines. Since spatio-temporal heterogeneity widely exists in practical datasets, the proposed framework may also cast light on other spatial-temporal applications. Model implementation is available at https://github.com/Echo-Ji/ST-SSL.

As a core technology of Intelligent Transportation System, traffic flow prediction has a wide range of applications. The fundamental challenge in traffic flow prediction is to effectively model the complex spatial-temporal dependencies in traffic data. Spatial-temporal Graph Neural Network (GNN) models have emerged as one of the most promising methods to solve this problem. However, GNN-based models have three major limitations for traffic prediction: i) Most methods model spatial dependencies in a static manner, which limits the ability to learn dynamic urban traffic patterns; ii) Most methods only consider short-range spatial information and are unable to capture long-range spatial dependencies; iii) These methods ignore the fact that the propagation of traffic conditions between locations has a time delay in traffic systems. To this end, we propose a novel Propagation Delay-aware dynamic long-range transFormer, namely PDFormer, for accurate traffic flow prediction. Specifically, we design a spatial self-attention module to capture the dynamic spatial dependencies. Then, two graph masking matrices are introduced to highlight spatial dependencies from short- and long-range views. Moreover, a traffic delay-aware feature transformation module is proposed to empower PDFormer with the capability of explicitly modeling the time delay of spatial information propagation. Extensive experimental results on six real-world public traffic datasets show that our method can not only achieve state-of-the-art performance but also exhibit competitive computational efficiency. Moreover, we visualize the learned spatial-temporal attention map to make our model highly interpretable.

Simulating the human mobility and generating large-scale trajectories are of great use in many real-world applications, such as urban planning, epidemic spreading analysis, and geographic privacy protect. Although many previous works have studied the problem of trajectory generation, the continuity of the generated trajectories has been neglected, which makes these methods useless for practical urban simulation scenarios. To solve this problem, we propose a novel two-stage generative adversarial framework to generate the continuous trajectory on the road network, namely TS-TrajGen, which efficiently integrates prior domain knowledge of human mobility with model-free learning paradigm. Specifically, we build the generator under the human mobility hypothesis of the A* algorithm to learn the human mobility behavior. For the discriminator, we combine the sequential reward with the mobility yaw reward to enhance the effectiveness of the generator. Finally, we propose a novel two-stage generation process to overcome the weak point of the existing stochastic generation process. Extensive experiments on two real-world datasets and two case studies demonstrate that our framework yields significant improvements over the state-of-the-art methods.

The recent success of large language models (LLMs) has shown great potential to develop more powerful conversational recommender systems (CRSs), which rely on natural language conversations to satisfy user needs.In this paper, we embark on an investigation into the utilization of ChatGPT for CRSs, revealing the inadequacy of the existing evaluation protocol.It might overemphasize the matching with ground-truth items annotated by humans while neglecting the interactive nature of CRSs.To overcome the limitation, we further propose an interactive Evaluation approach based on LLMs, named iEvaLM, which harnesses LLM-based user simulators.Our evaluation approach can simulate various system-user interaction scenarios.Through the experiments on two public CRS datasets, we demonstrate notable improvements compared to the prevailing evaluation protocol.Furthermore, we emphasize the evaluation of explainability, and Chat-GPT showcases persuasive explanation generation for its recommendations.Our study contributes to a deeper comprehension of the untapped potential of LLMs for CRSs and provides a more flexible and realistic evaluation approach for future research about LLMbased CRSs.The code is available at https: //github.com/RUCAIBox/iEvaLM-CRS.

Trajectory Representation Learning (TRL) is a powerful tool for spatial-temporal data analysis and management. TRL aims to convert complicated raw trajectories into low-dimensional representation vectors, which can be applied to various downstream tasks, such as trajectory classification, clustering, and similarity computation. Existing TRL works usually treat trajectories as ordinary sequence data, while some important spatial-temporal characteristics, such as temporal regularities and travel semantics, are not fully exploited. To fill this gap, we propose a novel Self-supervised trajectory representation learning framework with TemporAl Regularities and Travel semantics, namely START. The proposed method consists of two stages. The first stage is a Trajectory Pattern-Enhanced Graph Attention Network (TPE-GAT), which converts the road network features and travel semantics into representation vectors of road segments. The second stage is a Time-Aware Trajectory Encoder (TAT-Enc), which encodes representation vectors of road segments in the same trajectory as a trajectory representation vector, meanwhile incorporating temporal regularities with the trajectory representation. Moreover, we also design two self-supervised tasks, i.e., span-masked trajectory recovery and trajectory contrastive learning, to introduce spatial-temporal characteristics of trajectories into the training process of our START framework. The effectiveness of the proposed method is verified by extensive experiments on two large-scale real-world datasets for three downstream tasks. The experiments also demonstrate that our method can be transferred across different cities to adapt heterogeneous trajectory datasets.

Given its broad applications, time series analysis has gained substantial research attention but remains a very challenging task. Recent years have witnessed the great success of deep learning methods, eg., CNN and RNN, in time series classification and forecasting, but heterogeneity as the very nature of time series has not yet been addressed adequately and remains the performance "treadstone." In this light, we argue that the intra-sequence non-stationarity and inter-sequence asynchronism are two types of heterogeneities widely existed in multiple times series, and propose a hybrid attention network called WHEN as deep learning solution. WHEN features in two attention mechanisms in two different modules. In the WaveAtt module, we propose a novel data-dependent wavelet function and integrate it into the BiLSTM network as the wavelet attention, for the purpose of analyzing dynamic frequency components in nonstationary time series. In the DTWAtt module, we transform the dynamic time warping (DTW) technique into the form as the DTW attention, where all input sequences are synchronized with a universal parameter sequence to overcome the time distortion problem in multiple time series. WHEN with the hybrid attentions is then formulated as task-dependent neural network for either classification or forecasting tasks. Extensive experiments on 30 UEA datasets and 3 real-world datasets with rich competitive baselines demonstrate the excellent performance of our model. The ability of WHEN in dealing with time series heterogeneities is also elaborately explored via specially designed analysis.

In recent years, ride-on-demand (RoD) services such as Uber and DiDi are becoming increasingly popular. Different from traditional taxi services, RoD services adopt dynamic pricing mechanisms to manipulate the supply and demand on the road, and such mechanisms improve service capacity and quality. Seeking route recommendation has been widely studied in taxi service. In RoD service, the dynamic price is a new and accurate indicator describing the supply and demand, but it is yet rarely studied in providing clues for drivers to seek for passengers. In this paper, we propose to incorporate the impacts of dynamic prices as a key factor in recommending seeking routes to drivers. We first justfiy why it is necessary to recommend seeking routes and consider dynamic prices, by analyzing real service data from a typical RoD service. We then design a reinforcement learning model based on order and GPS trajectories datasets, and take into account dynamic prices in the design. Results prove that our model improves both driver earnings and seeking strategies. On driver earnings, the reinforcement learning model increases revenue efficiency by up to 34.52%, and considering dynamic prices leads to another increase of 6.19%. On seeking strategies, drivers are encouraged to serve local demand first, and they are redistributed more evenly and effectively.

Multivariate time series anomaly detection has great potentials in many practical applications. Extreme unbalanced training set and noise interference make it challenging to accurately capture the distribution of normal data and then detect anomalies. Existing AutoEncoder(AE)-based approaches are lack of effective regularization method specially designed for anomaly detection tasks thus easily overfitting while Generative Adversarial Network(GAN)-based approaches are mostly trained under the hypothesis of pollution-free training set, which means the training set is all composed of normal samples and that is hard to satisfy in practice. To tackle these problems, in this paper we propose a GAN based anomaly detection method for multivariate time series named FGANomaly (letter F is for Filter). The core idea is to filter possible anomalous samples with pseudo-labels before training the discriminator thus to capture the distribution of normal data as precise as possible. In addition, we design a novel training objective for the generator, which leads the generator to concentrate more on plausible normal data and ignore anomalies. We conducted comprehensive experiments on four public datasets, and the experimental results show the superiority of our method over baselines in both performance and robustness.

Traffic flow prediction is a fundamental problem in spatiotemporal data mining. Most of the existing studies focuses on designing statistical models to fit historical traffic data, which are purely data-driven approaches and fail to reveal the underlying mechanisms of urban traffic. To address this issue, we propose the spatiotemporal potential energy field model (ST-PEF+), which applies the field theory for human mobility to interpret the underlying mechanisms of urban traffic, and introduces the theory into data-driven deep learning models. ST-PEF+ consists of a PEF extraction module and a data-driven module. Inspired by the field theory for human mobility, the PEF extraction module adopts an algorithm to decompose the grid-based traffic flow graph into several polytree-based potential energy fields (PEFs), where traffic flows from high potential locations to low potential locations, just as water is driven by the gravity field. We also provide a theoretical analysis to ensure that the polytree decomposition algorithm can decompose any traffic flow graph. In the data-driven module, ST-PEF+ learns a spatiotemporal deep learning model to predict the dynamics of PEFs. The model adopts correlation-adaptive neural network structures, which consists of a temporal component for temporal correlations and a spatial component for spatial correlations. The temporal component employs a GRU and DCN combined structure to capture both short-term autocorrelation and long-term repeating patterns of PEFs. The spatial component extends the GAT using weighted directed attention to model the asymmetric spatial structure in PEFs. The prediction results of traffic flow are finally derived from PEFs that are predicted by the spatiotemporal deep learning model. We conduct extensive evaluations on three real-world traffic datasets. The results show that our model outperforms the state-of-the-art baselines. In addition, case studies confirm that the PEFs learned in our framework can reveal the underlying mechanisms of urban traffic, thus improving the model interpretability.

As travel efficiency matters to the work productivity of cities, shortening passengers' travel time for metros is therefore a pressing need. To this end, we study a strategy by dynamically scheduling dwell time for trains. Developing such a strategy is challenging because of three aspects: 1) Optimizing the average travel time of passengers needs to properly balance passengers' waiting time at platforms and journey time on trains, as well as considering long-term impacts; 2) Capturing dynamic spatio-temporal (ST) correlations of incoming passengers for metro stations is difficult; and 3) For each train, the dwell time scheduling is affected by other trains, which is hard to measure. To tackle these challenges, we propose a novel deep neural network, entitled AutoDwell. Specifically, AutoDwell optimizes the long-term rewards of dwell time settings in terms of passengers' waiting and journey time by a reinforcement learning framework. Next, AutoDwell employs gated recurrent units and graph attention networks to extract the ST correlations of the passenger flows among metro stations. Moreover, attention mechanisms are leveraged in AutoDwell for capturing the interactions between the trains. Extensive experiments on two real-world datasets demonstrate the superior performance of AutoDwell over several baselines, capable of saving passengers' travel time significantly.

To characterize complex and heterogeneous side information in recommender systems, the heterogeneous information network (HIN) has shown superior performance and attracted much research attention. In HIN, the rich entities, relations, and paths can be utilized to model the correlations of users and items; such a task setting is often calledHIN-based recommendation. Although HIN provides a general approach to modeling rich side information, it lacks special consideration on the goal of the recommendation task. The aggregated context from the heterogeneous graph is likely to incorporate irrelevant information, and the learned representations are not specifically optimized according to the recommendation task. Therefore, there is a need to rethink how to leverage the useful information from HIN to accomplish the recommendation task. To address the above issues, we propose a Curriculum pre-training based HEterogeneous Subgraph Transformer (calledCHEST) with newdata characterization,representation model,andlearning algorithm. Specifically, we consider extracting useful information from HIN to compose the interaction-specific heterogeneous subgraph, containing highly relevant context information for recommendation. Then, we capture the rich semantics (e.g., graph structure and path semantics) within the subgraph via a heterogeneous subgraph Transformer, where we encode the subgraph into multi-slot sequence representations. Besides, we design a curriculum pre-training strategy to provide an elementary-to-advanced learning process. The elementary course focuses on capturing local context information within the subgraph, and the advanced course aims to learn global context information. In this way, we gradually capture useful semantic information from HIN for modeling user-item interactions. Extensive experiments conducted on four real-world datasets demonstrate the superiority of our proposed method over a number of competitive baselines, especially when only limited training data is available.

High-performance traffic flow prediction model designing, a core technology of Intelligent Transportation System, is a long-standing but still challenging task for industrial and academic communities. The lack of integration between physical principles and data-driven models is an important reason for limiting the development of this field. In the literature, physics-based methods can usually provide a clear interpretation of the dynamic process of traffic flow systems but are with limited accuracy, while data-driven methods, especially deep learning with black-box structures, can achieve improved performance but can not be fully trusted due to lack of a reasonable physical basis. To bridge the gap between purely data-driven and physics-driven approaches, we propose a physics-guided deep learning model named Spatio-Temporal Differential Equation Network (STDEN), which casts the physical mechanism of traffic flow dynamics into a deep neural network framework. Specifically, we assume the traffic flow on road networks is driven by a latent potential energy field (like water flows are driven by the gravity field), and model the spatio-temporal dynamic process of the potential energy field as a differential equation network. STDEN absorbs both the performance advantage of data-driven models and the interpretability of physics-based models, so is named a physics-guided prediction model. Experiments on three real-world traffic datasets in Beijing show that our model outperforms state-of-the-art baselines by a significant margin. A case study further verifies that STDEN can capture the mechanism of urban traffic and generate accurate predictions with physical meaning. The proposed framework of differential equation network modeling may also cast light on other similar applications.

The availability of electronic health record data makes it possible to develop automatic disease diagnosis approaches. In this paper, we study the early diagnosis of diseases. As being a difficult task (even for experienced doctors), early diagnosis of diseases poses several challenges that are not well solved by prior studies, including insufficient training data, dynamic and complex signs of complications and trade-off between earliness and accuracy.

With increased human mobility and the introduction of NPIs, the complex, dynamic spread of COVID-19 has diverged significantly from SEIR's single, static assumption. At the same time, the ability to obtain front-line data also limits the modeling capabilities of SEIR. For researchers who cannot program, they must find suitable collaborators to implement their research. Even for researchers who can program, they need to repeat the principle and application process of the infectious disease model. LibEpidemic provide an open-source framework for modeling infectious disease, especially COVID-19, with bigdata. Researchers can implement subdivided, multi-stage or even metapopulation with the support of LibEpidemic.

With the unprecedented development of industrialization and urbanization, many hazardous chemicals have become an indispensable part of our daily life. They are produced, transported, and consumed in modern cities every day, which breeds many unknown hazardous chemicals-related locations (HCLs) that are out of the supervision of management departments and accompanying huge threats to urban safety. How to recognize these unknown HCLs and identify their risk levels is an essential task for urban hazardous chemicals management. To accomplish this task, in this work, we propose a system named as CityShield to discover hidden HCLs and classify their risk levels based on trajectories of hazardous chemicals transportation vehicles. The CityShield system consists of three components. The first component is Data Pre-processing, which filters noises in raw trajectories and probes stable transportation vehicles' stay points from massive uncertain GPS points. The second is HCL Recognition, which adopts the proposed HCL-Rec algorithm to cluster stay points into polygonal HCLs, and avoids the improper location merging problem caused by the skewed spatial distribution of HCLs. The third component is HCL Classification, which introduces the HCL relation graph as auxiliary information to overcome the label scarcity problem of HCLs. It adopts a self-supervised method consisting of four pre-training tasks to learn high-quality representations for HCLs from the graph, which are finally used to classify the categories and risk levels of HCLs.

In recent years, automatic computational systems based on deep learning are widely used in medical fields, such as automatic diagnosing and disease prediction. Most of these systems are designed for data sufficient scenarios. However, due to the disease rarity or privacy, the medical data are always insufficient. When applying these data-hungry deep learning models with insufficient data, it is likely to lead to issues of over-fitting and cause serious performance problems. Many data augmentation methods have been proposed to solve the data insufficiency problem, such as using GAN (Generative Adversarial Networks) to generate training data. However, the augmented data usually contains lots of noise. Directly using them to train sensitive medical models is very difficult to achieve satisfactory results.

Junyi Li, Tianyi Tang, Zheng Gong, Lixin Yang, Zhuohao Yu, Zhipeng Chen, Jingyuan Wang, Xin Zhao, Ji-Rong Wen. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022.

Limited by the statistical-based machine learning framework, a spurious correlation is likely to appear in existing knowledge-aware recommendation methods. It refers to a knowledge fact that appears causal to the user behaviors (inferred by the recommender) but is not in fact. For tackling this issue, we present a novel approach to discovering and alleviating the potential spurious correlations from a counterfactual perspective. To be specific, our approach consists of two counterfactual generators and a recommender. The counterfactual generators are designed to generate counterfactual interactions via reinforcement learning, while the recommender is implemented with two different graph neural networks to aggregate the information from KG and user-item interactions respectively. The counterfactual generators and recommender are integrated in a mutually collaborative way. With this approach, the recommender helps the counterfactual generators better identify potential spurious correlations and generate high-quality counterfactual interactions, while the counterfactual generators help the recommender weaken the influence of the potential spurious correlations simultaneously. Extensive experiments on three real-world datasets have shown the effectiveness of the proposed approach by comparing it with a number of competitive baselines. Our implementation code is available at: https://github.com/RUCAIBox/CGKR.

In this work, we study an important task in location-based services, namelyPersonalized Route Recommendation (PRR). Given a road network, the PRR task aims to generate user-specific route suggestions for replying to users’ route queries. A classic approach is to adapt search algorithms to construct pathfinding-like solutions. These methods typically focus on reducing search space with suitable heuristic strategies. For these search algorithms, heuristic strategies are often handcrafted, which are not flexible to work in complicated task settings. In addition, it is difficult to utilize useful context information in the search procedure. To develop a more principled solution to the PRR task, we propose to improve search algorithms with neural networks for solving the PRR task based on the widely used$A^{*}$algorithm. The main idea of our solution is to automatically learn the cost functions in$A^{*}$algorithms, which is the key of heuristic search algorithms. Our model consists of two main components. First, we employ attention-based Recurrent Neural Networks (RNN) to model the cost from the source to the candidate location by incorporating useful context information. Instead of learning a single cost value, the RNN component is able to learn a time-varying vectorized representation for the moving state of a user. Second, we propose to use an estimation network for predicting the cost from a candidate location to the destination. For capturing structural characteristics, the estimation network is built on top of position-aware graph attention networks. The two components are integrated in a principled way for deriving a more accurate cost of a candidate location for the$A^{*}$algorithm. Extensive experiment results on three real-world datasets have shown the effectiveness and robustness of the proposed model.

The rapidly-growing business of ride-on-demand (RoD) service such as Uber, Lyft and Didi proves the effectiveness of their new service model – using mobile apps and dynamic pricing to coordinate between drivers, passengers and the service provider, to manipulate the supply and demand, and to improve service responsiveness as well as quality. Despite its success, dynamic pricing creates a new problem for drivers: how to seek for passengers to maximize revenue under dynamic prices. Seeking route recommendation has already been studied extensively in traditional taxi service, but most studies do not consider the effects of taxis and passengers on the seeking taxi simultaneously. Further, in RoD service it is necessary to consider more factors such as dynamic prices, the status of other transportation services, etc. In this paper, we employ a force-directed approach to model, by analogy, the relationship between vacant cars and passengers as that between positive and negative charges in electrostatic field. We extract features from multi-source urban data to describe dynamic prices, the status of RoD, taxi and public transportation services, and incorporate them into our model. The model is then used in route recommendation in every intersection so that a driver in a vacant RoD car knows which road segment to take next. We conduct extensive experiments based on our multi-source urban data, including RoD service operational data, taxi GPS trajectory data and public transportation distribution data, and results not only show that our approach outperforms existing baselines, but also justify the need to incorporate multi-source urban data and dynamic prices.

With the increase of traffic prediction models, there has become an urgent need to develop a standardized framework to implement and evaluate these methods. This paper presents LibCity, a unified, comprehensive, and extensible library for traffic prediction, which provides researchers with a credible experimental tool and a convenient development framework. In this library, we reproduce 42 traffic prediction models and collect 29 spatial-temporal datasets, which allows researchers to conduct comprehensive experiments in a convenient way. To accelerate the development of new models, we design unified model interfaces based on unified data formats, which effectively encapsulate the details of the implementation. To verify the effectiveness of our implementations, we also report the reproducibility comparison results of LibCity, and set up a performance leaderboard for the four kinds of traffic prediction tasks. Our library will contribute to the standardization and reproducibility in the field of traffic prediction. The open source link of LibCity is https://github.com/LibCity/Bigscity-LibCity.

With the development of electronic health records (EHRs), prenatal care examination records have become available for developing automatic prediction or diagnosis approaches with machine learning methods. In this paper, we study how to effectively learn representations applied to various downstream tasks for EHR data. Although several methods have been proposed in this direction, they usually adapt classic sequential models to solve one specific diagnosis task or address unique EHR data issues. This makes it difficult to reuse these existing methods for the early diagnosis of pregnancy complications or provide a general solution to address the series of health problems caused by pregnancy complications. In this paper, we propose a novel model RAPT, which stands for RepresentAtion by Pre-training time-aware Transformer. To associate pre-training and EHR data, we design an architecture that is suitable for both modeling EHR data and pre-training, namely time-aware Transformer. To handle various characteristics in EHR data, such as insufficiency, we carefully devise three pre-training tasks to handle data insufficiency, data incompleteness and short sequence problems, namely similarity prediction, masked prediction and reasonability check. In this way, our representations can capture various EHR data characteristics. Extensive experimental results for four downstream tasks have shown the effectiveness of the proposed approach. We also introduce sensitivity analysis to interpret the model and design an interface to show results and interpretation for doctors. Finally, we implement a diagnosis system for pregnancy complications based on our pre-training model. Doctors and pregnant women can benefit from the diagnosis system in early diagnosis of pregnancy complications.

The recent pandemic has painfully reminded us of the need for IS design to be informed by possible futures and conscious of undesirable futures – Within months, many of the nice-to-have IS functionalities have become must-haves; Technology solutions in response to the pandemic have raised privacy and equality concerns. Although design science research fundamentally focuses on shaping artefacts and events to create a more desirable future, there has been limited guidance on how futures should be accounted for. This article addresses the gap by integrating insights from future-oriented IS research and futures research to develop guidelines for engaging with futures throughout the design science research process. The future-ready design research (FRDR) process prompts researchers to be more aware of futures, to foster the innovative foresight for actively pursuing the preferred future, and to espouse the responsible foresight for consciously avoiding undesirable futures. The guidelines are illustrated with a design science research project on outbreak analytics and the instantiated system’s subsequent adaptation and utilisation in COVID-19.

The fuzzy cognitive map (FCM) is a powerful model for system state prediction and interpretable knowledge representation. Recent years have witnessed the tremendous efforts devoted to enhancing the basic FCM, such as introducing temporal factors, uncertainty or fuzzy rules to improve interpretation, and introducing fuzzy neural networks or wavelets to improve time series prediction. But how to achieve high-precision yet interpretable prediction in cross-domain real-life applications remains a great challenge. In this article, we propose a novel FCM extension called deep FCM (DFCM) for multivariate time series forecasting, in order to take both the advantage of FCM in interpretation and the advantage of deep neural networks in prediction. Specifically, to improve the predictive power, DFCM leverages a fully connected neural network to model connections (relationships) among concepts in a system, and a recurrent neural network to model unknown exogenous factors that have influences on system dynamics. Moreover, to foster model interpretability encumbered by the embedded deep structures, a partial derivative-based approach is proposed to measure the connection strengths between concepts in DFCM. An alternate function gradient descent algorithm is then proposed for parameter inference. The effectiveness of DFCM is validated over four publicly available datasets with the presence of seven baselines. DFCM indeed provides an important clue to building interpretable predictors for real-life applications.

With the development of location-acquisition technologies, there are a huge number of mobile trajectories generated and accumulated in a variety of domains. However, due to the constraints of device and environment, many trajectories are recorded at low sampling rate, which increases the uncertainty between two consecutive sampled points in the trajectories. Our task is to recover a high-sampled trajectory based on the irregular low-sampled trajectory in free space, i.e., without road network information. There are two major problems with traditional solutions. First, many of these methods rely on heuristic search algorithms or simple probabilistic models. They cannot well capture complex sequential dependencies or global data correlations. Second, for reducing the predictive complexity of the unconstrained numerical coordinates, most of the previous studies have adopted a common preprocessing strategy by mapping the space into discrete units. As a side effect, using discrete units is likely to bring noise or inaccurate information. Hence, a principled post-calibration step is required to produce accurate results, which has been seldom studied by existing methods. To address the above difficulties, we propose a novel Deep Hybrid Trajectory Recovery model, named DHTR. Our recovery model extends the classic sequence-to-sequence generation framework by implementing a subsequence-to-sequence recovery model tailored for the current task, named subseq2seq. In order to effectively capture spatiotemporal correlations, we adopt both spatial and temporal attentions for enhancing the model performance. With the attention mechanisms, our model is able to characterize long-range correlations among trajectory points. Furthermore, we integrate the subseq2seq with a calibration component of Kalman filter (KF) for reducing the predictive uncertainty. At each timestep, the noisy predictions from the subseq2seq component will be fed into the KF component for calibration, and then the refined predictions will be forwarded to the subseq2seq component for the computation of the next timestep. Extensive results on real-world datasets have shown the superiority of the proposed model in both performance and interpretability.

Recent years have witnessed the emergence of worldwide megalopolises and the accompanying public safety events, making urban safety a top priority in modern urban management. Among various threats, dangerous goods such as gas and hazardous chemicals transported through cities have bred repeated tragedies and become the deadly “bomb” we sleep with every day. While tremendous research efforts have been devoted to dealing with dangerous goods transportation (DGT) issues, further study is still in great need to quantify this problem and explore its intrinsic dynamics from a big data perspective. In this article, we present a novel system called DGeye , to feature a fusion between DGT trajectory data and residential population data for dangers perception and prediction. Specifically, DGeye first develops a probabilistic graphical model-based approach to mine spatio-temporally adjacent risk patterns from population-aware risk trajectories. Then, DGeye builds the novel causality network among risk patterns for risk pain-point identification, risk source attribution, and online risky state prediction. Experiments on both Beijing and Tianjin cities demonstrate the effectiveness of DGeye in real-life DGT risk management. As a case in point, our report powered by DGeye successfully drove the government to lay down gas pipelines for the famous Guijie food street in Beijing.

Traffic flow prediction is of great importance in traffic management and public safety, but is challenging due to the complex spatial-temporal dependencies as well as temporal dynamics. Existing work either focuses on traditional statistical models, which have limited prediction accuracy, or relies on black-box deep learning models, which have superior prediction accuracy but are hard to interpret. In contrast, we propose a novel interpretable spatiotemporal deep learning model for traffic flow prediction. Our main idea is to model the physics of traffic flow through a number of latent Spatio-Temporal Potential Energy Fields (ST-PEFs), similar to water flow driven by the gravity field. We develop a Wind field Decomposition (WD) algorithm to decompose traffic flow into poly-tree components so that ST-PEFs can be established. We then design a spatiotemporal deep learning model for the ST-PEFs, which consists of a temporal component (modeling the temporal correlation) and a spatial component (modeling the spatial dependencies). To the best of our knowledge, this is the first work that make traffic flow prediction based on ST-PEFs. Experimental results on real-world traffic datasets show the effectiveness of our model compared to the existing methods. A case study confirms our model interpretability.

While having achieved great success in rich real-life applications, deep neural network (DNN) models have long been criticized for their vulnerability to adversarial attacks. Tremendous research efforts have been dedicated to mitigating the threats of adversarial attacks, but the essential trait of adversarial examples is not yet clear, and most existing methods are yet vulnerable to hybrid attacks and suffer from counterattacks. In light of this, in this paper, we first reveal a gradient-based correlation between sensitivity analysis-based DNN interpreters and the generation process of adversarial examples, which indicates the Achilles's heel of adversarial attacks and sheds light on linking together the two long-standing challenges of DNN: fragility and unexplainability. We then propose an interpreter-based ensemble framework called X-Ensemble for robust adversary defense. X-Ensemble adopts a novel detection-rectification process and features in building multiple sub-detectors and a rectifier upon various types of interpretation information toward target classifiers. Moreover, X-Ensemble employs the Random Forests (RF) model to combine sub-detectors into an ensemble detector for adversarial hybrid attacks defense. The non-differentiable property of RF further makes it a precious choice against the counterattack of adversaries. Extensive experiments under various types of state-of-the-art attacks and diverse attack scenarios demonstrate the advantages of X-Ensemble to competitive baseline methods.

Road network is the core component of urban transportation, and it is widely useful in various traffic-related systems and applications. Due to its important role, it is essential to develop general, effective, and robust road network representation models. Although several efforts have been made in this direction, they cannot fully capture the complex characteristics of road networks.

Recent years have witnessed the world-wide emergence of mega-metropolises with incredibly huge populations. Understanding residents mobility patterns, or urban dynamics, thus becomes crucial for building modern smart cities. In this paper, we propose a Neighbor-Regularized and context-aware Non-negative Tensor Factorization model (NR-cNTF) to discover interpretable urban dynamics from urban heterogeneous data. Different from many existing studies concerned with prediction tasks via tensor completion, NR-cNTF focuses on gaining urban managerial insights from spatial, temporal, and spatio-temporal patterns. This is enabled by high-quality Tucker factorizations regularized by both POI-based urban contexts and geographically neighboring relations. NR-cNTF is also capable of unveiling long-term evolutions of urban dynamics via a pipeline initialization approach. We apply NR-cNTF to a real-life data set containing rich taxi GPS trajectories and POI records of Beijing. The results indicate: 1) NR-cNTF accurately captures four kinds of city rhythms and seventeen spatial communities; 2) the rapid development of Beijing, epitomized by the CBD area, indeed intensifies the job-housing imbalance; 3) the southern areas with recent government investments have shown more healthy development tendency. Finally, NR-cNTF is compared with some baselines on traffic prediction, which further justifies the importance of urban contexts awareness and neighboring regulations.

Recent years have witnessed the rapidly-growing business of ride-on-demand (RoD) services such as Uber, Lyft and Didi. Unlike taxi services, these emerging transportation services use dynamic pricing to manipulate the supply and demand, and to improve service responsiveness and quality. Despite this, on the drivers' side, dynamic pricing creates a new problem: how to seek for passengers in order to earn more under the new pricing scheme. Seeking strategies have been studied extensively in traditional taxi service, but in RoD service such studies are still rare and require the consideration of more factors such as dynamic prices, the status of other transportation services, etc. In this paper, we develop ROD-Revenue, aiming to mine the relationship between driver revenue and factors relevant to seeking strategies, and to predict driver revenue given features extracted from multi-source urban data. We extract basic features from multiple datasets, including RoD service, taxi service, POI information, and the availability of public transportation services, and then construct composite features from basic features in a product-form. The desired relationship is learned from a linear regression model with basic features and high-dimensional composite features. The linear model is chosen for its interpretability-to quantitatively explain the desired relationship. Finally, we evaluate our model by predicting drivers' revenue. We hope that ROD-Revenue not only serves as an initial analysis of seeking strategies in RoD service, but also helps increasing drivers' revenue by offering useful guidance.

The deep network model, with the majority built on neural networks, has been proved to be a powerful framework to represent complex data for high performance machine learning. In recent years, more and more studies turn to nonneural network approaches to build diverse deep structures, and the Deep Stacking Network (DSN) model is one of such approaches that uses stacked easy-to-learn blocks to build a parameter-training-parallelizable deep network. In this paper, we propose a novel SVM-based Deep Stacking Network (SVM-DSN), which uses the DSN architecture to organize linear SVM classifiers for deep learning. A BP-like layer tuning scheme is also proposed to ensure holistic and local optimizations of stacked SVMs simultaneously. Some good math properties of SVM, such as the convex optimization, is introduced into the DSN framework by our model. From a global view, SVM-DSN can iteratively extract data representations layer by layer as a deep neural network but with parallelizability, and from a local view, each stacked SVM can converge to its optimal solution and obtain the support vectors, which compared with neural networks could lead to interesting improvements in anti-saturation and interpretability. Experimental results on both image and text data sets demonstrate the excellent performances of SVM-DSN compared with some competitive benchmark models.

Fastest Route Recommendation (FRR) aims to find the fastest path in response to user's queries in a large complex road network. Early studies cast the FRR task as a pathfinding problem on graphs and adopt heuristic algorithms as the major solution due to the efficiency and robustness. A major problem of heuristic algorithms is that the heuristic function is usually empirically set with simple methods, which is difficult to model other useful factors. In this paper, we extend the classic A* algorithm for the FRR task by modeling complex traffic information with neural networks. Specially, we identify an important factor that is important to improve the FRR task, i.e. the estimation of travel time. For this purpose, we first develop a module for predicting the time-varying traffic speed for a road segment, which is the foundation for estimating the travel time. Conditioned on this module, we further design another module to estimate the fastest travel time between two locations connected by routes. We adopt neural networks to implement both modules for enabling the capacity of modeling complex traffic characteristics and dynamics. In this way, the original two cost functions of A* algorithm have been set in a more principled way with neural networks. To our knowledge, we are the first to use neural networks for improving A* algorithm in the FRR task. It elegantly combines the merits of A* algorithm and the powerful modeling capacities of neural networks for the FRR task. Extensive results on the three real-world datasets have shown the effectiveness and robustness of the proposed model.

Personalized Route Recommendation (PRR) aims to generate user-specific route suggestions in response to users' route queries. Early studies cast the PRR task as a pathfinding problem on graphs, and adopt adapted search algorithms by integrating heuristic strategies. Although these methods are effective to some extent, they require setting the cost functions with heuristics. In addition, it is difficult to utilize useful context information in the search procedure. To address these issues, we propose using neural networks to automatically learn the cost functions of a classic heuristic algorithm, namely A* algorithm, for the PRR task. Our model consists of two components. First, we employ attention-based Recurrent Neural Networks (RNN) to model the cost from the source to the candidate location by incorporating useful context information. Instead of learning a single cost value, the RNN component is able to learn a time-varying vectorized representation for the moving state of a user. Second, we propose to use a value network for estimating the cost from a candidate location to the destination. For capturing structural characteristics, the value network is built on top of improved graph attention networks by incorporating the moving state of a user and other context information. The two components are integrated in a principled way for deriving a more accurate cost of a candidate location. Extensive experiment results on three real-world datasets have shown the effectiveness and robustness of the proposed model.

Recent years have witnessed the successful marriage of finance innovations and AI techniques in various finance applications including quantitative trading (QT). Despite great research efforts devoted to leveraging deep learning (DL) methods for building better QT strategies, existing studies still face serious challenges especially from the side of finance, such as the balance of risk and return, the resistance to extreme loss, and the interpretability of strategies, which limit the application of DL-based strategies in real-life financial markets. In this work, we propose AlphaStock, a novel reinforcement learning (RL) based investment strategy enhanced by interpretable deep attention networks, to address the above challenges. Our main contributions are summarized as follows: i) We integrate deep attention networks with a Sharpe ratio-oriented reinforcement learning framework to achieve a risk-return balanced investment strategy; ii) We suggest modeling interrelationships among assets to avoid selection bias and develop a cross-asset attention mechanism; iii) To our best knowledge, this work is among the first to offer an interpretable investment strategy using deep reinforcement learning models. The experiments on long-periodic U.S. and Chinese markets demonstrate the effectiveness and robustness of AlphaStock over diverse market states. It turns out that AlphaStock tends to select the stocks as winners with high long-term growth, low volatility, high intrinsic value, and being undervalued recently.

Driven by the wave of urbanization in recent decades, the research topic about migrant behavior analysis draws great attention from both academia and the government. Nevertheless, subject to the cost of data collection and the lack of modeling methods, most of existing studies use only questionnaire surveys with sparse samples and non-individual level statistical data to achieve coarse-grained studies of migrant behaviors. In this paper, a partially supervised cross-domain deep learning model named CD-CNN is proposed for migrant/native recognition using mobile phone signaling data as behavioral features and questionnaire survey data as incomplete labels. Specifically, CD-CNN features in decomposing the mobile data into location domain and communication domain, and adopts a joint learning framework that combines two convolutional neural networks with a feature balancing scheme. Moreover, CD-CNN employs a three-step algorithm for training, in which the co-training step is of great value to partially supervised cross-domain learning. Comparative experiments on the city Wuxi demonstrate the high predictive power of CD-CNN. Two interesting applications further highlight the ability of CD-CNN for in-depth migrant behavioral analysis.

Recent years have witnessed the unprecedented rising of time series from almost all kindes of academic and industrial fields. Various types of deep neural network models have been introduced to time series analysis, but the important frequency information is yet lack of effective modeling. In light of this, in this paper we propose a wavelet-based neural network structure called multilevel Wavelet Decomposition Network (mWDN) for building frequency-aware deep learning models for time series analysis. mWDN preserves the advantage of multilevel discrete wavelet decomposition in frequency learning while enables the fine-tuning of all parameters under a deep neural network framework. Based on mWDN, we further propose two deep learning models called Residual Classification Flow (RCF) and multi-frequecy Long Short-Term Memory (mLSTM) for time series classification and forecasting, respectively. The two models take all or partial mWDN decomposed sub-series in different frequencies as input, and resort to the back propagation algorithm to learn all the parameters globally, which enables seamless embedding of wavelet-based frequency analysis into deep learning frameworks. Extensive experiments on 40 UCR datasets and a real-world user volume dataset demonstrate the excellent performance of our time series models based on mWDN. In particular, we propose an importance analysis method to mWDN based models, which successfully identifies those time-series elements and mWDN layers that are crucially important to time series analysis. This indeed indicates the interpretability advantage of mWDN, and can be viewed as an indepth exploration to interpretable deep learning.

Since the 21st century, the global outbreaks of infectious diseases such as SARS in 2003, H1N1 in 2009, and H7N9 in 2013, have become the critical threat to the public health and a hunting nightmare to the government. Understanding the propagation in large-scale metapopulations and predicting the future outbreaks thus become crucially important for epidemic control and prevention. In the literature, there have been a bulk of studies on modeling intra-city epidemic propagation but with the single population assumption (homogeneity). Some recent works on metapopulation propagation, however, focus on finding specific human mobility physical networks to approximate diseases transmission networks, whose generality to fit different diseases cannot be guaranteed. In this paper, we argue that the intra-city epidemic propagation should be modeled on a metapopulation base, and propose a two-step method for this purpose. The first step is to understand the propagation system by inferring the underlying disease infection network. To this end, we propose a novel network inference model called D 2 PRI, which reduces the individual network into a sub-population network without information loss, and incorporates the power-law distribution prior and data prior for better performance. The second step is to predict the disease propagation by extending the classic SIR model to a metapopulation SIR model that allows visitors transmission between any two sub-populations. The validity of our model is testified on a real-life clinical report data set about the airborne disease in the Shenzhen city, China. The D 2 PRI model with the extended SIR model exhibit superior performance in various tasks including network inference, infection prediction and outbreaks simulation.

Ride-on-demand (RoD) services such as Uber and Didi (in China) are becoming increasingly popular, and in these services dynamic price plays an important role in balancing the supply (i.e., the number of cars) and demand (i.e., the number of passenger requests) to benefit both drivers and passengers. However, the dynamic price also creates concerns for passengers: the "unpredictable" prices sometimes prevent them from making quick decisions at ease. One may wonder if it is possible to get a lower price if s/he chooses to wait a while. Giving passengers more information helps to tackle this concern, and predicting the prices is a possible solution.

Customer volume prediction, which predicts the volume from a customer source to a service place, is a very important technique for location selection, market investigation, and other related applications. Most of traditional methods only make use of partial information for either supervised or unsupervised modeling, which cannot well integrate overall available knowledge. In this paper, we propose a method titled GR-NMF for jointly modeling both implicit correlations hidden inside customer volumes and explicit geographical knowledge via an integrated probabilistic framework. The effectiveness of GR-NMF in coupling all-round knowledge is verified over a real-life outpatient dataset under different scenarios. GR-NMF shows particularly evident advantages to all baselines in location selection with the cold-start challenge.

Recent years have witnessed the continuous growth of megalopolises worldwide, which makes urban safety a top priority in modern city life. Among various threats, dangerous goods such as gas and hazardous chemicals transported through and around cities have increasingly become the deadly "bomb" we sleep with every day. In both academia and government, tremendous efforts have been dedicated to dealing with dangerous goods transportation (DGT) issues, but further study is still in great need to quantify the problem and explore its intrinsic dynamics in a big data perspective. In this paper, we present a novel system called DGeye, which features a "duet" between DGT trajectory data and human mobility data for risky zones identification. Moreover, DGeye innovatively takes risky patterns as the keystones in DGT management, and builds causality networks among them for pain point identification, attribution and prediction. Experiments on both Beijing and Tianjin cities demonstrate the effectiveness of DGeye. In particular, the report generated by DGeye driven the Beijing government to lay down gas pipelines for the famous Guijie food street.

The project named after City-Eyes is based on the HBASE big data analyzing platform and the real-time-display technology with H5/CSS3, which could analyze data and result dynamically. The project is devoted to designing the online city-data-analyzing platform, with interacting in real time, filtering, analyzing and displaying data. In the view of filtering, we sample Wuxi and collect much information, including phone signaling, railway, airports, Weibo, population mobility and etc. Moreover, we write enough data mining scripts for varieties of data to extract the value of them. In the view of analyzing, we build an efficient analyzing model, which divides the task into two parts, which includes the front end and the back end. The back end is focusing on compressing data and extracting valid information by clustering the data streams. At the same time, the front-end applies the real-time-display technology to provide an interactive and nice-looking data visualizing solution for our users. In the view of displaying, we come up with various solutions from two angles for every variety of data: 'distribution and network'. With svg/canvas animation tech, the data can be mixed into the map and display itself dynamically.

Traffic speed prediction is a long-standing and critically important topic in the area of Intelligent Transportation Systems (ITS). Recent years have witnessed the encouraging potentials of deep neural networks for real-life applications of various domains. Traffic speed prediction, however, is still in its initial stage without making full use of spatio-temporal traffic information. In light of this, in this paper, we propose a deep learning method with an Error-feedback Recurrent Convolutional Neural Network structure (eRCNN) for continuous traffic speed prediction. By integrating the spatio-temporal traffic speeds of contiguous road segments as an input matrix, eRCNN explicitly leverages the implicit correlations among nearby segments to improve the predictive accuracy. By further introducing separate error feedback neurons to the recurrent layer, eRCNN learns from prediction errors so as to meet predictive challenges rising from abrupt traffic events such as morning peaks and traffic accidents. Extensive experiments on real-life speed data of taxis running on the 2nd and 3rd ring roads of Beijing city demonstrate the strong predictive power of eRCNN in comparison to some state-of-the-art competitors. The necessity of weight pre-training using a transfer learning notion has also been testified. More interestingly, we design a novel influence function based on the deep learning model, and showcase how to leverage it to recognize the congestion sources of the ring roads in Beijing.

The Transport Control Protocol (TCP) has been widely used by wired and wireless Internet applications such as FTP, email and HTTP. Numerous congestion algorithms have been proposed to improve the performance of TCP in various scenarios, especially for high bandwidth-delay product (BDP) and wireless networks. Although different algorithms may achieve different performance improvements under different network conditions, designing a congestion algorithm that performs well across a wide spectrum of network conditions remains a great challenge. In this paper, we propose a novel congestion control algorithm, named TCP-FIT, which could perform gracefully in both wireless and high BDP networks. The algorithm was inspired by parallel TCP, but with the important distinctions that only one TCP connection with one congestion window is established for each TCP session, and that no modifications to other layers (e.g. the application layer) of the end-to-end system need to be made. Extensive experimental results obtained using both network simulators as well as over “live” wired line, WiFi and 3G networks at different geographical locations and at different times of the day are presented. The performance of the algorithm shown in the experiment results is significantly improved as compared to other state-of-the-art algorithms, while maintaining good fairness.

In this paper, we propose a novel scheme using computing complexity layered scalable video coding (CCLSVC) to optimize the user experience of broadcasting video in the computing capability limited handheld terminals. To address the heterogeneity of computing capability among different handheld devices, we employ hierarchal B reference structure of SVC to divide the frames into multiple computing complexity layers (CC Layers) in server side. The handheld clients simply choose to decode the frames in their corresponding layers in terms of their computation capability to maximize the video PSNR. We have proved that the optimal CC Layers division problem is a precedence constrained scheduling problem, which is an NP-complete problem. And we further propose our fast greedy method to approximately get optimized broadcasting video playback PSNR. The simulation shows that our method is superior to temporal SVC and random frame discarding method.